We have seen the results of testing. M2 appears to basically be m1.1. It is still based on 5nm and the speed increase appears to be directly related to the increase in cores/die size.
If anything this makes me appreciative of my m1 MacBook Air as it is holding strong performance wise this year. I will definitely be waiting for the 3nm m3 as I believe the performance and battery efficiency will be sog if I a toy better.
Why do you all think?
This sort of thinking is based on the CPU being the only important part of a SoC. That may be true for your use cases, but it's not so for everyone.
While the CPUs on the M2 are reasonably characterized as you say, this is not true of the GPU.
It's probably not true of the NPU (though we don't yet have benchmarks).
It may well not be true of AMX (the patent trail suggests that AMX is evolving very rapidly to something like a competitor for AVX512, though it's never clear when a patent translates into a shipping product).
Even on the media side, it's probably the case that the M2 supports h.266 (at least decode), so that (just like Apple handled h.265) when they support the switchover to .266 as the preferred codec, there will already be a large installed base.
Bottom line is, right now all we have is quick benchmarks (like GB5) for CPU and some limited aspects of GPU. Until we get the full exploration (which, for M1, took many months) it's way too early to say.